TITLE:
Problems with Double Blind in Medicine
AUTHORS:
Yngvar Reichelt, Karl L. Reichelt
KEYWORDS:
Diagnosis, Difference, Etiologies, Distribution, Groups
JOURNAL NAME:
Open Journal of Psychiatry,
Vol.5 No.2,
April
22,
2015
ABSTRACT: Double blind tests of
drugs and procedures depend on obtaining two equal and randomly assigned groups
to be compared. With a diagnosis based only on symptoms, but with different
etiologies, this is not very likely. We here show the probability of obtaining
two equal groups with one diagnosis but three etiologies. The mathematical name
for such a problem is multivariate hypergeometric distribution. We find that
increasing the group size decreases the probability.